Tweet
I'm having a little trouble believing the RAR column on the pitcher projections. To take an illustrative example,
Al Lieter is projected for a 3.45 ERA in 203 IPs; Eric Milton is projected at a 3.72 ERA in 173 IPs. Lieter is projected to pitch better and for more innings than Milton, so it follows that his RAR should be higher, right? And yet, Milton is listed higher in RAR at 23.5 vs Lieter's 21.5.
A quick glance at these discrepancies hints that perhaps you are using xERA to calculate RAR instead of ERA, but I think that just shifts the question then: why are some pitchers projected to have higher ERA than xERA and vice versa? If there is a known causal factor that differentiates between the two of them, then shouldn't you add that causal factor into xERA and thus create a better formula? And if there isn't, then is it anything but random guessing to project gaps one way for some people and another way for others?
BTW, these small differences don't make all that much difference for auction leagues, but they can make a pretty big difference for draft leagues, where *ordering* determines your draft pick... a little nudge here or there to a projection and all the sudden John Thomson looks like a high round draft choice, when I really don't think he should be. I've read Ron's column in which he says he creates his projections to "nudge" us in certain directions, but I think that's somewhat dangerous in a drafting vs an auction context (for the record, this is one of the few philosophical problems I've ever had with something Ron has said: if I wanted to be manipulated with inaccurate numbers, I'd get my projections from a politician or, oh, my ex-wife. I don't want to be *manipulated*; I want to be given as many facts as possible and then treated with enough respect to allow me to decide what to do with those facts. I'm smart enough to read an honest projection for a possible breakout candidate and make a decision to "go the extra dollar" without those projections being artificially inflated to have me do so).
Anyway, I just want to make sure I believe these RAR numbers: these types of numbers are helpful for Scoresheet players.
Andy
Al Lieter is projected for a 3.45 ERA in 203 IPs; Eric Milton is projected at a 3.72 ERA in 173 IPs. Lieter is projected to pitch better and for more innings than Milton, so it follows that his RAR should be higher, right? And yet, Milton is listed higher in RAR at 23.5 vs Lieter's 21.5.
A quick glance at these discrepancies hints that perhaps you are using xERA to calculate RAR instead of ERA, but I think that just shifts the question then: why are some pitchers projected to have higher ERA than xERA and vice versa? If there is a known causal factor that differentiates between the two of them, then shouldn't you add that causal factor into xERA and thus create a better formula? And if there isn't, then is it anything but random guessing to project gaps one way for some people and another way for others?
BTW, these small differences don't make all that much difference for auction leagues, but they can make a pretty big difference for draft leagues, where *ordering* determines your draft pick... a little nudge here or there to a projection and all the sudden John Thomson looks like a high round draft choice, when I really don't think he should be. I've read Ron's column in which he says he creates his projections to "nudge" us in certain directions, but I think that's somewhat dangerous in a drafting vs an auction context (for the record, this is one of the few philosophical problems I've ever had with something Ron has said: if I wanted to be manipulated with inaccurate numbers, I'd get my projections from a politician or, oh, my ex-wife. I don't want to be *manipulated*; I want to be given as many facts as possible and then treated with enough respect to allow me to decide what to do with those facts. I'm smart enough to read an honest projection for a possible breakout candidate and make a decision to "go the extra dollar" without those projections being artificially inflated to have me do so).
Anyway, I just want to make sure I believe these RAR numbers: these types of numbers are helpful for Scoresheet players.
Andy
Comment